For reprint information contact:
Belhhan Akpinar, MD Tel: 90 212 224 4950 Fax: 90 212 239 8791 email: belh@turk.net
Department of Cardiovascular Surgery, Kadir Has University Medical Faculty, Florence Nightingale Hospital, Abide-I Hürriyet Caddesi, No. 290 S¸is¸li, Istanbul, Turkey.
A CUTE G ASTROINTESTINAL
C OMPLICATIONS A FTER O PEN
H EART S URGERY
Belhhan Akpinar, MD, Ertan Sa˘gbas¸, MD, Mustafa Güden, MD, Kubilay Kemertas¸, MD
1, Bingür Sönmez, MD, Osman Bayindir, MD
2, Cem’i Demirog˘lu, MD
3Department of Cardiovascular Surgery
1
Department of General Surgery
2
Department of Anesthesia
3
Department of Cardiology
Kadir Has University Medical Faculty Florence Nightingale Hospital
Istanbul, Turkey
A
BSTRACTRetrospective analysis revealed that 24 of 4401 adult patients (0.5%) developed severe gastrointestinal complications after open heart surgery during a 3-year period from January 1995. There were 4 women (17%) and 20 men (83%). Mean age was 61.7 ± 2.02 years. Gastrointestinal bleeding (33.3%), mesenteric ischemia (20.8%), pancreatitis (20.8%), hepatic dysfunction (16.7%), and cholecystitis (16.7%) were the most common complications. Mortality was 41.7% (10 patients).
During the same period, mortality in the patients who did not develop gastrointestinal complications was 1.89% (p < 0.0001). Emergency basis, reoperation, combined operations, peripheral vascular disease, diabetes mellitus, chronic lung disease, and impaired left ventricle function were found to be risk factors for the development of postoperative gastrointestinal complications.
(Asian Cardiovasc Thorac Ann 2000;8:109–13)
INTRODUCTION
Gastrointestinal (GI) complications affect approximately 1% to 2% of patients undergoing open heart surgery.1,2 Although this is a low incidence, such complications can increase morbidity, length of postoperative hospital stay, and mortality.3,4 Since the onset is early in the post- operative period when the patient is fully sedated, identi- fication may be difficult because symptoms may be masked by postoperative analgesia. This retrospective study reports our experience with acute GI complications after cardiac surgery during a 3-year period.
PATIENTS AND METHODS
All 4401 adult patients who underwent open heart surgery
in our institution between January 1995 and January 1998 were retrospectively evaluated for early major and minor GI complications (within 7 days after operation). A major GI complication was defined as one that required consultation with an abdominal surgeon, transfer to an intensive care unit, surgical intervention, or treatment that prolonged the hospital stay.
Preoperatively, all patients were premedicated with midazolam 0.07 to 0.1 mg·kg–1 intramuscularly. Anesthesia was induced with fentanyl 25 µg·kg–1 and midazolam 0.15 to 0.35 mg·kg–1 with vecuronium 0.14 to 0.18 mg·kg–1 for muscle relaxation. Anesthesia was maintained with 60% nitrous oxide in oxygen and supplementary doses of
fentanyl and midazolam. All patients received standard doses of morphine 10 to 20 mg·kg–1·h–1 and midazolam infusion at 8 to 15 µg·kg–1·h–1 for basal sedation and analgesia in the early postoperative period. Diltiazem infusion at 0.1 mg·kg–1·h–1 was started at the induction of anesthesia and continued in the intensive care unit for 24 hours.
Myocardial protection during cardiopulmonary bypass (CPB) and aortic crossclamping was achieved by infusion of antegrade-retrograde tepid (32°C) blood cardioplegia.
The perfusate was allowed to cool to 32°C and all patients received nonpulsatile flow at rates of 2.2 to 2.4 L·min–1·m–2 to maintain a mean perfusion pressure of 55 to 60 mm Hg. Epinephrine infusion at 0.01 to 0.1 µg·kg–1·min–1 was the first choice of inotropic agent during weaning from CPB, regulated according to hemodynamic para- meters. Dobutamine was added as a second inotropic drug when necessary, instead of raising epinephrine doses.
Cefuroxime was given for antibiotic prophylaxis, H2- receptor antagonists were used routinely for prophylaxis against GI bleeding.
The unpaired Student’s t test and the chi-squared test with the Yates correction were employed to assess differences between means and groups.
RESULTS
Twenty-four patients (0.5%) developed major GI compli- cations after cardiac surgery. Patient characteristics are shown in Table 1. Operative procedures in the 24 patients with major GI complications are listed in Table 2.
Perioperative variables that were analyzed in both groups are given in Table 3.
Thirty major GI complications were diagnosed in the 24 patients; the type, mode of treatment, and mortality are
shown in Table 4. Bleeding occurred in 8 patients (33.3%) between the 1st and 4th postoperative days (mean, 52 ± 6.2 hours). Diagnosis was made by endoscopy in 5 cases, clinical judgment was used in the other 3. Bleeding was stopped by nasogastric tube irrigation with cold saline and antacid solution in 3 patients, one underwent a successful endoscopic procedure to coagulate the bleeding area, and 4 with unremitting hemorrhage underwent surgery. Two of these also had duodenal perforation and died of multiorgan failure. This group of patients needed massive blood transfusions (mean, 5.4 ± 2 units). The 5 patients with mesenteric ischemia underwent reoperation;
4 were still intubated when the diagnosis was made.
Mean onset of symptoms was 30 ± 4.4 hours. Severe distension, undefined metabolic acidosis, hemodynamic instability, and leukocytosis (> 30 × 109/L) were common findings. Although bedside ultrasonography was per- formed in most patients, diagnosis was confirmed by bedside minilaparotomy. One patient who could undergo localized bowel resection survived but the other 4 had necrosis too extensive for resection; one of these also had splanchnic necrosis. One patient who developed signs of peritonitis on the second postoperative day underwent a minilaparotomy that revealed a perforated diverticulum of the colon. The diverticula was sutured primarily and a temporary colostomy was performed. The colostomy was closed 3 months later. Four patients were diagnosed with acute cholecystitis, by ultrasonography in 2 and by exploratory laparotomy in the other 2. Three of these patients underwent successful open cholecystectomy, the other was controlled medically and surgery was performed one month later. Five patients developed pancreatitis that was subclinical in 3 who did well with medical treatment.
Two with fulminant pancreatitis died and postmortem examination revealed necrotizing pancreatitis. Hepatic dysfunction was defined as abnormal liver function tests,
Table 1. Profile of 4401 Open Heart Surgery Patients
Major GI Complications No or Minor GI Complications
Variable n % n % p
Total 24 0.5 4377 99.5
Mean age (years) 61.7 ± 2.02 60.2 ± 11.9 > 0.05
Male:Female ratio 5:1 4:1
Emergency operation 4 16.7 220 5 < 0.001
Routine operation 20 83.3
LVEF < 30% 6 25 525 12 < 0.05
Redo operations 3 12.5 92 2.1 < 0.00l
Combined operations 11 45.8 488 11.1 < 0.001
IABP 1 4.2 42 1 < 0.001
Diabetes mellitus 8 33.3 812 18.5 < 0.002
PVD 7 29.2 82 1.9 < 0.001
COPD 3 12.5 196 4.5 < 0.002
Renal dysfunction* 0 0 32 0.7 > 0.05
*Preoperative serum creatinine > 16 mg·L–1. COPD = chronic obstructive pulmonary disease, GI = gastrointestinal, IABP = intraaortic balloon pump, LVEF = left ventricular ejection fraction, PVD = peripheral vascular disease.
hyperbilirubinemia, elevated bleeding time, and decreased albumin levels. Two of the 4 patients with severe hepatic dysfunction died from multiorgan failure; postmortem examination in 1 revealed severe hepatocellular necrosis.
Preoperative liver function tests were normal in all 4 cases.
During the same period, 68 patients were diagnosed as having mild GI problems such as distension, decrease in bowel sounds, paralytic ileus, diarrhea and mild jaundice.
They were evaluated as a separate group because symptoms were not life-threatening, did not cause any organ dysfunction, and subsided in 2 to 5 days without the need for surgical intervention. GI distension developed in 48 patients, which was diagnosed as paralytic ileus.
They were kept on total parenteral nutrition with nasogastric tube decompression until the return of peristaltic activity. In some cases, a rectal tube was also used. One case of distension was diagnosed as colonic pseudo-obstruction and treated with colonoscopy.
Table 2. Cardiac Operations in 24 Patients With Major Gastrointestinal Complications
Operation No. of Patients
CABG 12
Valve reconstruction 1
CABG + valve replacement 4
CABG + LV aneurysm repair 3
CABG + PV surgery 4
CABG = coronary artery bypass grafting, LV = left ventricular, PV = peripheral vascular.
Nineteen of these patients had elevated liver enzymes and slight increases in bilirubin and amylase levels without any morbidity or mortality. Mortality in the 4377 patients who developed no or only mild GI complications was 1.89% compared with 41.6% in those with major GI complications (p < 0.0001).
Analysis of the cardiac surgery patients with major GI complications revealed that emergency basis, combined operations, left ventricular ejection fraction less than 30%, diabetes mellitus, peripheral vascular disease, and chronic obstructive lung disease were risk factors for early GI complications (Table 1).
DISCUSSION
The incidence of complications involving the entire GI tract after cardiac surgery has been reported as between 1% and 2% and associated mortality rates ranged from 40% to 60%.5–7 The common mechanism is a low cardiac output state that contributes to inadequate tissue per- fusion, mucosal ischemia, and “acute GI focal necrosis syndrome.” Stress ulceration, mucosal atrophy, bacterial overgrowth from stress-ulcer prophylaxis, and loss of barrier function with increased permeability may lead to bacterial translocation, sepsis, and multiorgan failure.6,8 The findings in this study agree with previous reports that advanced age, reoperation, combined or complex surgical procedures, and emergency surgery are risk factors for GI complications.2,4,8 In addition, impaired left ventricular function, peripheral vascular disease, and chronic obstructive lung disease also emerged as risk factors.
Low cardiac output requiring high doses of inotropics,
Table 4. Major Gastrointestinal Complications and Mortality in 24 Patients
Complication n % Deaths
Gastrointestinal bleeding 6 25 0
Gastrointestinal bleeding + duodenal perforation 2 8.3 2
Mesenteric ischemia 5 20.8 4
Pancreatitis 5 2.8 2
Hepatic dysfunction 4 16.7 2
Acute cholecystitis 4 16.7 0
Bowel diverticula perforation 1 4.2 0
Acute appendicitis 1 4.2 0
Table 3. Operative and Postoperative Variables
Major Gastrointestinal No or Minor Gastrointestinal
Variable Complications Complications p
Cardiopulmonary bypass time (min) 108.73±13.47 83.62±14.70 < 0.05
Crossclamp time (min) 64.50±9.64 41.20±8.50 < 0.05
New myocardial infarction (n) 4/24 (16.7%) 72/4377 (1.6%) < 0.002
Inotropics > 0.03 µg·kg–1 epinephrine (n) 12/24 (50%) 786/4377 (18%) < 0.002
Intensive care unit stay (days) 6.96±1.91 1.26±0.7 < 0.001
especially in patients with poor ventricular function, further aggravates acidosis, vasoconstriction, and GI dysfunction.
Although preoperative renal function was normal in those who developed GI complications, the incidence of postoperative renal dysfunction was much higher, mainly because of multiorgan failure.
The presence of fever, high white blood cell count, abdominal distension, marked tenderness on abdominal palpation, hemodynamic evidence of sepsis, or positive blood cultures can be manifestations of an acute intra- abdominal phenomenon. Reaching the correct diagnosis is usually very difficult. Laboratory tests may not dis- tinguish cholecystitis from perforated viscus, mesenteric ischemia, or pancreatitis. Bedside abdominal ultrasound and endoscopy are sometimes helpful but abdominal computed tomography and angiography are difficult to perform in these critically ill patients. Peritoneal lavage and exploratory minilaparotomy are most helpful for diagnosis.7,8 Endoscopy and bedside laparotomy pro- cedures were readily performed in this study.
Among 8 patients who had upper GI tract bleeding, 2 had a history of gastritis and one had a preexisting ulcer. They developed GI bleeding in spite of taking H2-receptor blockers. Patients who responded to medical treatment (3), or endoscopic laser bipolar coagulation (1) did well.
Four patients with unremitting hemorrhage had to undergo reoperation, of whom 2 died from duodenal perforation.
In view of the high mortality, prevention is most important and any patient with a history of ulcer should receive medication to prevent stress-related mucosal damage and GI bleeding.9–11 In addition, any patient on prolonged ventilatory support should receive stress-ulcer prophy- laxis.12–14
Mesenteric ischemia is usually noted in elderly patients who have prolonged low cardiac output states requiring pharmacological or mechanical support. Patients with atherosclerotic aorta are also at risk of mesenteric ischemia due to embolism.15,16 Profound ileus or abdominal pain out of proportion with physical findings, sepsis, acidosis, and respiratory distress are common findings.16 Although the diagnosis can be made by arteriography, this is usually not practical for a critically ill patient and a minilaparotomy is more convenient for diagnosis. Four of the 5 patients who underwent abdominal exploration for bowel necrosis died. The survivor had an early laparotomy and could undergo localized bowel resection, whereas the others had extensive necrosis and multiple areas of ischemic bowel that did not permit resection. Common laboratory findings in this group were leukocytosis and high levels of serum lactate.
Hyperamylasemia has been noted in up to 35% of patients following CPB, but it is associated with clinical pancreatitis in only 1% to 3%.17,18 Three patients in our series were diagnosed with subclinical pancreatitis. They
had mild symptoms of anorexia, nausea, and ileus, with elevation of serum lipase levels. A brief period of bowel rest was sufficient in these patients. On the other hand, two patients who developed fulminant pancreatitis died.
Postmortem examination in one of these patients revealed necrotizing pancreatitis.
There has been a documented shift towards increasing age and severity of illness in the patient population undergoing cardiac surgery who are prone to com- plications.4,8 Patients with GI complications usually have a very complex course after their cardiac surgery, such as cardiogenic shock, respiratory and renal failure prior to the development of their acute surgical problem.7 Older patients, especially those with a history of GI problems, prolonged CPB times, low cardiac output, or prolonged ventilatory support, should be under greater suspicion.
Surgery must be performed promptly because unnecessary delay is associated with unacceptable mortality rates.8 We believe that good collaboration and cooperation with a general surgery team experienced with postcardiotomy GI complications is essential. Liberal use of bedside exploratory small laparotomy and endoscopic techniques should be encouraged since inappropriate conservative therapy is poorly tolerated in this high-risk group. The frustratingly high mortality rates due to late diagnosis calls for a low threshold to exploratory minilaparotomy.
Prevention and early diagnosis seem to be the only reliable assets for better results. Only heightened awareness and prompt surgical action will succeed in reducing the high mortality.
REFERENCES
1. Tsiotos GG, Mullany CJ, Zietlow S, van Heerden JA.
Abdominal complication following cardiac surgery. Am J Surg 1994;167:553–7.
2. Ott MJ, Buchman TG, Baumgartner WA. Postoperative abdominal complications in cardiopulmonary bypass patients: a case-controlled study. Ann Thorac Surg 1995;59:1210–3.
3. Christenson JT, Schmuziger M, Maurice J, Simonet F, Velebit V. Gastrointestinal complications after coronary artery bypass grafting. J Thorac Cardiovasc Surg 1994;108:899–906.
4. Perugini RA, Orr RK, Porter D, Dumas EM, Maini BS.
Gastrointestinal complications following cardiac surgery.
An analysis of 1477 cardiac surgery patients. Arch Surg 1997;132:352–7.
5. Spotnitz WD, Sanders RP, Hanks JB, Nolan SP, Tribble CG, Berginj D. General surgical complications can be predicted after cardiopulmonary bypass. Ann Surg 1995;221:489–97.
6. Baue AE. The role of the gut in the development of multiple organ dysfunction in cardiothoracic patients. Ann Thorac Surg 1993;55:822–9.
7. Mercado PD, Farid H, O’Cannell TX, Sintek CF, Pfeffer T, Khonsari S. Gastrointestinal complications associated with cardiopulmonary bypass procedures. Am J Surg 1994;60:789–92.
8. Bojar RM. Manuel of perioperative care in cardiac sur- gery. 3rd ed. Malden: Blackwell Science, 1999:404–12.
9. Cook DJ, Fuller HD, Guyatt GH, Marshall JC, Leasa D, Hall R. Risk factors for gastrointestinal bleeding in critically ill patients. N Eng J Med 1994;330:377–81.
10. Rosen HR, Vlahakes GJ, Rattner DW. Fulminant peptic ulcer disease in cardiac surgical patients: pathogenesis, prevention, and management. Crit Care Med 1992;20:
354–9.
11. Eggleston JM, Pearson RC, Holland J, Tooth JA, Vohra A, Doran BH. Prospective endoscopic study of stress erosions and ulcers in critically ill adult patients treated with sucralfate or placebo. Crit Care Med 1994;22:1949–54.
12. Cook D, Guyatt G, Reeve BK, Heylend DK, Griffith E, Marshall J. Stress ulcer prophylaxis in critically ill patients.
Resolving discordant meta analyses. JAMA 1996;275:
308–14.
13. Philips JO, Metzler MH, Palmieri TL, Huckfeldt RE, Dahl NG. A prospective study of simplified omeprazole
suspension for the prophylaxis of stress-related mucosal damage. Crit Care Med 1996;24:1793–1800.
14. Rovers JP, Souney PF. A critical review of continuous infusion H2 receptor therapy. Crit Care Med 1989;17:
814–21.
15. Allen KB, Salam AA, Lumsden AB. Acute mesenteric ischemia after cardiopulmonary bypass. J Vasc Surg 1992;16:391–6.
16. Kaleya RN, Sammartano RJ, Boley SJ. Aggressive approach to acute mesenteric ischemia. Surg Clin N Am 1992;72:157–82.
17. Rattner DW, Gu ZY, Vlahakes GJ, Warshaw AL.
Hyperamylasemia after cardiac surgery. Incidence, significance, and management. Ann Surg 1989;209:
279–83.
18. Svensson LG, Decker G, Kinsley RB. A prospective study of hyperamylasemia and pancreatitis after cardiopulmo- nary bypass. Ann Thorac Surg 1985;39:409–11.
